4602 zones

This release has a zoning system on "Gemeindeverband"-level, counting 4602 zones. Transport networks are refined accordingly. The road network now also includes tertiary roads. Unless there is 4 TB RAM available, the resulting size of the OD table requires OD sampling. Each set of scenarios has a consistent OD set (inherited through the scenario hierarchy) that is sampled among major population centres, random zones, and a set of scenario-relevant zones specified in parameters.xls. Zone clustering and NUTS3-level modelling is still possible, but discuraged due to lower accuracy.

There are also two more demand segments than before, as the purpose "buy/execute" (formerly included for compatibility with "Verkehrsverflechtungsprognose 2030") is now split into "shopping" and "errands". Travel purposes now fully align with "Mobilität in Deutschland". All demand models have been re-estimated and refined accordingly. Compulsory trip purposes are distributed with a doubly constrained assignment, including inner/inter-zonal distinction. It performs well based on employment data and OSM-sourced education data. The destination choice model for non-compulsory trips performs slightly worse, but is required to connect volumes to POI data from OSM. Though, it is not possible to estimate a valid inner/inter-zonal choice model, even after many trials (it would also lack logic). The generation choice model, too, is logically not fully sound, which is why exogenous trip generation from MiD data is recommended (though it does not depict elasticity of demand).

Additionally, except some bugfixes and documentation clarifications, the following changes were applied:

• quetzal_germany now uses the library quetzal-lite, which is easy to install and compatible with new versions of pandas and other dependencies
• Generation volumes and inner/inter-zonal shares are exogenously computed from MiD on RegioStaR7-level
• Volumes for compulsory and non-compulsory trips are saved in different files, making the model structure more modular
• Distances in choice models are now euclidian distances rather than road network distances, making the model less error-prone
• The composite cost file is now in the input_static folder

Policies

This release enhances ease of use for policy scenarios, even though the model structure stays unchanged. The major novelty is a new tab in the parameters.xls that includes easy-to-understand policy levers. The following policy implementations are the main contributions:

• generic cycling highways depending on population density thresholds
• create a generic on-demand ride pooling network
• add search time for parking in urban areas
• connection to a car ownership model, runnable as an additional module, if desired
• reactivation of 215 closed rail lines across Germany in ample detail

There were also some clarifications in the Readme and the launcher, and bugfixes.

Sufficiency scenarios

This release creates three maximum sufficiency scenarios, incorporating a large set of political, technological, social, economic, and organisational drivers, attributed to the corresponding scenario: Avoid, Shift, Avoid+Shift (compared to a reference scenario). Results indicate which transport systems could be possible in Germany.

This release added some policy levers to the notebooks, accessible via the corresponding parameters in parameters.xls:

• Add car sharing to the mode choice model
• Change the average speed of rail_long services between major cities (urbanisation = 1; population > 100,000)
• Add car-free inner cities, together with park-and-ride option in mode choice model
• Influence employed persons, employment, and the number of POIs by urbanisation degree

Additionally, this release

• makes the full model runnable from the 00_launcher notebook
• enhances the documentation of modelling steps and their order
• refined the OD set sampling for computers with standard RAM
• added validation notebook category and a scenario validation dashboard
• fixed time issue for non-motorised paths and re-calibrated
• reworked the inner-zonal modelling (independent on exogenous data; i.e. all assumptions moved to parameters.xls)
• connects price sensitivity parameter to input parameter income change
• fixed some bugs

Full endogenous demand model

All steps of the demand model (generation, distribution, and mode choice) are now implemented endogenously without reliance on external data sources for the OD matrix. All discrete choice models are calibrated with MiD2017. Optionally, one can use the exogenous OD matrix from VP2030, still.

The demand model structure is as follows: The mode choice logit model provides OD-specific composite cost (CC), which feed into the destination choice logit model (MNL) for non-obligatory trip purposes (buy/execute, leisure, accompany). The decision of going beyond the home cell's boundaries is subject to another binomial logit model, which takes the zone's mean CC as input, amongst other variables. Trip generation for these purposes is an MNL with the number of trips per day per person (0 to 5) for the corresponding purpose, also informed by mean CC of the corresponding origin zone.
Compulsory trip purposes (commuting, business, education) are computed with a doubly constrained distribution, using inverted CC as deterrence matrix.

Other major developments of this release are:

1. The mode choice model considers disaggregated public transport modes (coach and bus, as well as long- and short-distance rail services).
2. PT subscriptions are an exogenous parameter.
3. Sparsification of the OD set for smaller, more efficient models to be ran on laptops. The OD sample size can be defined as scenario parameter.
4. Employment, employed persons data and the quantity of various points of interest are included on zone level
5. Policy measures:
   5.1. Local on-demand transport in non-urban areas
   5.2. Ban of air transport under a distance threshold
   5.3. Car traffic speed limits for motorways, inner-urban, and the rest of the roads
6. Automation bugfixes

2225 zones

A finer zoning system yields more detailed results in aggregated transport models. Model zones are now based on "Gemeindeverband"-level within Germany and agglomaratively clustered in oder to yield similar zone sizes and a viable model size. The old NUTS3-level zoning system can still be applied easily, as described in the 00_launcher notebook. Results are as valid as in the previous release.

The finer model requires around 150GB RAM when the LoS tables are used (i.e. in all modelling steps). Moreover, the new version of input_static must be downloaded from

Car availability segmentation

This release adds demand segments for car availability to quetzal_germany. All previous demand segments (i.e. trip purposes) are further differentiated into the availability of at least one car in the household, as travel behaviour varies significantly between these user groups. This differentiation allows more advanced travel behaviour and policy measure analysis. The share of persons with car(s) available is a suitable interface for car ownership models.

All notebooks now take demand segments as defined in the parameters.xls settings. Results of the mode choice model are as valid as in v1.0.0.
Bugfix: Notebooks don't produce assertion errors when executed from the launcher, unless required for model results.

First release

This release incorporates

• a detailed, realistic network model for the region of Germany based on open data,
• level-of-service attributes time and price for seven different modes,
• 401 traffic zones on NUTS3-level resolution with corresponding socio-demographic data,
• an accurate mode choice model based on MiD2017 B2 data,
• calculation of traffic by mode within traffic zones,
• calculation of emissions for the entire transport system.

Compared to official statistics, quetzal_germany yields valid results for mode choice between traffic zones (except lower estimations for air traffic). It does not yet include endogenous simulation of trip volumes (currently coming from the "Verkehrsverflechtungsprognose 2030"), which produces inconsistencies in the distance distribution of trips. Resulting pkm show an over-estimation of car travel and an under-estimation of long-distance rail travel within reasonable bounds. Results can be used for any susequent calculations, which do not explicitly focus on air traffic. Find them attached for pkm in the trip's origin zone on NUTS3-level and on NUTS1-level with additional information for emissions, travel time, and prices.